Garbage Collecting the World Wide Web

Stephen Watt

Abstract:   The World Wide Web has grown over the past decade and a
half from being a collection of modest resources to an essential part
of our every day lives.   While the curation of early web sites could
be managed by hand, present web portals are backed by collections of
numerous hosts with complex inter-connections.   The issues arising in
the management of present-day web resources bears many similarities to
the issues encountered with explicit storage management in the first
few decades with programming languages.   For example, the owner of a
web site may remove a page when it is still referenced by others, or a
page may be kept long after it is no longer accessible.      Lisp, a
programming language for symbolic computation, introduced the notion
of garbage collection for automatic storage management.   We show how
the ideas of modern distributed garbage collection can be applied to
the world wide web and present the results of some experiments.